chore: release v1.6.0

[StuMason]

Summary

Release v1.6.0 with 8 new database creation tools.

Added

• Database Creation Tools - Full CRUD support for all database types:
  • create_postgresql - Create PostgreSQL databases
  • create_mysql - Create MySQL databases
  • create_mariadb - Create MariaDB databases
  • create_mongodb - Create MongoDB databases
  • create_redis - Create Redis databases
  • create_keydb - Create KeyDB databases
  • create_clickhouse - Create ClickHouse databases
  • create_dragonfly - Create Dragonfly databases (Redis-compatible)

Changed

• Total tool count increased from 67 to 75 tools
• Database tools section now has 14 tools (was 6)

Test plan

• All 201 tests pass
• Build succeeds
• Lint passes
• Documentation updated (README, CHANGELOG, CLAUDE.md)

🤖 Generated with Claude Code

[claude]

PR Review: v1.6.0 - Database Creation Tools

Overall Assessment

This is a well-structured PR that adds 8 new database creation tools following the established codebase patterns. The implementation is clean and consistent with existing conventions.

---

✅ Strengths

Code Quality

• Excellent use of a shared databaseBaseSchema object to reduce code duplication across the 8 database creation tools
• TypeScript interfaces properly extend a base interface (CreateDatabaseBaseRequest)
• Client methods have explicit return types as required by the project's TypeScript standards
• Consistent naming conventions across all new tools

Architecture

• Properly follows the 4-step pattern outlined in CLAUDE.md: types → client → mcp-server → tests
• API endpoints match Coolify's REST API structure (/databases/{type})
• Clean separation between client layer and MCP tool definitions

Documentation

• CHANGELOG follows Keep a Changelog format
• README and CLAUDE.md updated with new tool counts
• Database category in feature table properly updated

---

⚠️ Issues Found

1. Tool Count Discrepancy (Minor)
The PR documentation states 75 tools, but counting this.tool( calls in mcp-server.ts shows 77 tools. The documentation should be updated to reflect the accurate count.

2. MCP Server Tests Only Test Mock Functions Directly
The tests in mcp-server.test.ts (lines 396-484) call the mock functions directly rather than testing through the MCP server's tool handlers:

// Current approach - tests the mock, not the integration
await mockCreatePostgresql({ ...baseParams, postgres_user: 'myuser' });
expect(mockCreatePostgresql).toHaveBeenCalledWith({...});

This pattern differs from other tests in the file that use spies on the actual client:

// Better approach used elsewhere
const spy = jest.spyOn(server['client'], 'getResource').mockResolvedValue({...});
await server.get_resource('test-uuid');
expect(spy).toHaveBeenCalledWith('test-uuid');

The current tests verify the mock is callable but don't actually exercise the MCP tool → client integration. Consider updating to match the spy pattern used in coolify-client.test.ts.

---

🔒 Security Considerations

• Password fields are appropriately handled as optional strings
• No hardcoded defaults for sensitive fields like passwords
• is_public option is properly documented to make users aware they're exposing the database publicly

---

📊 Test Coverage

• 8 client tests for database creation ✅
• 8 MCP server tests ✅ (though see note above about test pattern)
• All 201 tests pass per PR description

---

💡 Suggestions (Non-blocking)

1. Consider adding input validation for public_port (valid port range 1-65535)
2. File header comment in mcp-server.ts (line 10) says "Databases: list, get, start/stop/restart" but now includes create/delete operations - could be updated for accuracy

---

Verdict

Approve with minor fixes recommended. The core implementation is solid and follows project conventions. The tool count discrepancy in documentation should be corrected before merge.